How can I resample nonuniform data with sample and hold method?

Hi everyone. I am new to resampling, and am getting a little lost while trying to implement the sample and hold method.

I have a matrix of 4 columns, and varying number of rows of data (and the gaps are filled by NaNs).

eg:R= [1 2 3 4 5; 1 2 NaN NaN 6; 1 2 NaN NaN 9; NaN 3 NaN NaN 5; NaN 4 NaN NaN NaN];

I have a second matrix of time series, having same structure as the above matrix. Each non-NaN cell in R has a corresponding timestamp.

I want to upsample R, such that all the columns have 100 entries, but I do not want to interpolate between values. What I want to do is use the 'nearest' function in interp1 or n=0 in resample, but not sure how to do it with this data structure. The attempts I made either give me an error of 'not enough inputs', or give me a resampled vector with extra zeros in it. I am not sure how to do this.

Basically, let's say my timestamps start at -1 s and ends at 20 s. For every point in time that there is not recorded value, I want it to take the value of the previous closest value, until it comes to the next value.

eg:(showing just one column here):

time=[-1 2 3 4 6 8 10 12 14 19] 
R=[1 4 4 6 1 3 2 9 32 16];

What_I_want=[1 1 1 4 4 6 6 1 1 3 3 2 2 9 9 32 32 32 32 32 16 16];%so there is values for -1s to 20s, my fs here is 1Hz, if I understand correctly.

Any and all help will be very helpful. I have tried to explain as well as I could. I am pretty new to this. If any further information will be useful, please let me know. I realize I have been staring at my data for so long, that I may take some inherent features for granted. Hope people can understand my query.

Answer 1

[HMidas]

Hi, I figured it out. I just had to use the 'previous' option in interp1, not 'nearest'.

Thanks to everyone who read this :)